We study how well the complex gas velocity fields induced by massive spiral arms are modelled by the hydrodynamical simulations we used to constrain the dark matter fraction in nearby spiral galaxies (Kranz et al. 2001, 2003). More specifically, we explore the dependence of the positions and amplitudes of features in the gas flow on the temperature of the interstellar medium (assumed to behave as a one-component isothermal fluid), the non-axisymmetric disk contribution to the galactic potential, the pattern speed, Omega_p, and finally the numerical resolution of the simulation. We argue that, after constraining the pattern speed reasonably well by matching the simulations to the observed spiral arm morphology, the amplitude of the non-axisymmetric perturbation (the disk fraction) is left as the primary parameter determining the gas dynamics. However, due to the sensitivity of the positions of the shocks to modeling parameters, one has to be cautious when quantitatively comparing the simulations to observations. In particular, we show that a global least squares analysis is not the optimal method for distinguishing different models as it tends to slightly favor low disk fraction models. Nevertheless, we conclude that, given observational data of reasonably high spatial resolution and an accurate shock-resolving hydro-code this method tightly constrains the dark matter content within spiral galaxies. We further argue that even if the perturbations induced by spiral arms are weaker than those of strong bars, they are better suited for this kind of analysis because the spiral arms extend to larger radii where effects like inflows due to numerical viscosity and morphological dependence on gas sound speed are less of a concern than they are in the centers of disks.